In two blog posts, FTC Chief Technologist Ed Felten discusses the efficacy, or inefficacy, of hashing and pseudonyms when it comes to data anonymization. Hashing is a mathematical function often used to create pseudonyms for sensitive data such as Social Security numbers, for example. But Felten says the assumption that hashing "is sufficient to anonymize data is risky at best, and usually wrong." In addition, he writes, an online pseudonym can be tracked over time, revealing details about a user that could eventually lead to identification. The practice of using Social Security numbers as pseudonyms is risky, Felten writes, because they're shared across many data collectors for decades and are difficult to change.
Full Story